search

meeg-ml-benchmarks / brain-age-benchmark-paper

M/EEG brain age benchmark paper
https://meeg-ml-benchmarks.github.io/brain-age-benchmark-paper/
BSD 3-Clause "New" or "Revised" License
24 stars 10 forks source link

compute_benchmark_age_prediction.py line 315 error #52

Closed anthonyromyn closed 2 years ago

anthonyromyn commented 2 years ago

Hey brain-age benchmark team,

I've encountered an error that I'm having trouble solving. Running the shallow benchmark analysis on the chbp data, I'm getting the error in the picture below after cross-validation is done.

Screenshot (1332)

line 315 ys.loc[cv_splits[:, 1], 'cv_split'] = cv_splits[:, 0].astype(int)

ys and cv_splits are different lengths, so cv_splits[:,0] has many more rows relative to ys, thus the code fails to find appropriate places for many of the cv_splits

ys length equals the number of subjects cv_splits length is ~3221

Handcrafted, dummy, filterbank-riemann have all worked well. Haven't tried deep yet.

Thanks, Anthony

anthonyromyn commented 2 years ago

Can now confirm the same happens with the deep model

agramfort commented 2 years ago

@robintibor @gemeinl @hubertjb would you have time to look? and see if you can replicate eg in the compute system in Freiburg?

anthonyromyn commented 2 years ago

Thanks @agramfort and all!

hubertjb commented 2 years ago

@anthonyromyn Thanks for reporting this issue! #53 should fix this.

anthonyromyn commented 2 years ago

Awesome thanks @hubertjb, just ran a quick test and looks to have worked wonderfully, thanks!